ADVANCES IN MATHEMATICAL PHYSICS, vol.2022, 2022 (SCI-Expanded)
More recently, type-3 (T3) fuzzy logic systems (FLSs) with better learning ability and uncertainty modeling have been presented. On other hand, the proportional-integral-derivative (PID) is commonly employed in most industrial control systems, because of its simplicity and efficiency. The measurement errors, nonlinearities, and uncertainties degrade the performance of conventional PIDs. In this study, for the first time, a new T3-FLS-based PID scheme with deep learning approach is introduced. In addition to rules, the parameters of fuzzy sets are also tuned such that a fast regulation efficiency is obtained. Unlike the most conventional approaches, the suggested tuning approach is done in an online scheme. Also, a nonsingleton fuzzification is suggested to reduce the effect of sensor errors. The proposed scheme is examined on a case-study microgrid (MG), and its good frequency stabilization performance is demonstrated in various hard conditions such as variable load, unknown dynamics, and variation in renewable energy (RE) sources.